Logic can be a powerful tool for helping non-expert users cope with the complexity of the law. It can become a particularly powerful legal problem diagnosis tool.
It’s helpful to work through an example using a typical piece of legislation. Let’s assume our user has a specific type of legal issue under the fictional Widgets Act.
Let’s say the Widgets Act has 250 active sections. If we include all the subsections, it amounts to 1000 individual provisions.
To perform a legal diagnosis, we have to find the 1 relevant provision out of 1000 that applies to our user’s problem.
For this example, assume that sections 1-124 deal with white widgets and 125-250 deal with black widgets. Our user sees this question:
Our user inputs her information into the system to confirm she has a black widget. With one question, we know our user’s problem is addressed by sections 125-250 of the Widgets Act.
We’re down to roughly 125 sections or 500 provisions.
Now let’s ask another question. Half of the remaining provisions deal with round black widgets. The other provisions deal with square black widgets.
Our user sees this question:
Our user inputs her information into the system to confirm her black widget is square. With two questions, we know our user’s problem is addressed by roughly 62 sections of the Widgets Actor approximately 250 provisions.
See the pattern? With three questions, we are down to roughly 125 provisions out of 1000.
In four questions, we’re down to 63 provisions.
In 5 questions, we’re down to roughly 32.
In 6 questions, we’re down to 16.
In 7 questions, we’re down to 8.
In 8 questions, we’re down to 4.
In 9 questions, we’re down to 2 provisions.
In 10 questions, we’re down to 1 provision out of 1000 provisions in the Widgets Act. This is the provision that applies to our user’s problem.
Using a logic-based approach, it took us 10 questions to take our user to this conclusion.
Statutes aren’t so easy to divide the way we did in our example. But expert system questions can actually be structured to work faster (e.g. to require fewer questions) to narrow subject matter.
Our example only used an A / B structure that would have 2 possible answers at a time. Many expert systems would allow us to use 8 or 1o answers at at time. Imagine if the input possibilities looked like this:
An input structure that uses multiple answers can make the logic more powerful and efficient.
In only 3 or 4 questions, an expert system could diagnose a non-expert user’s legal problem down to 1 in 1000 and diagnose the user’s particular problem. The logic would apply expert reasoning to cut through complexity and improve access to justice.